[egs] multi_cn: fix false removal of 1st field of clean text (kaldi-asr#4020)

Author: naxingyu

egs/multi_cn/s5/local/train_lms.sh

@@ -29,6 +29,7 @@ fi
 
 cleantext=$dir/text.no_oov
 
+# note: ignore 1st field of text, it's the utterance-id.
 cat $text | awk '{$1=""; print substr($0, 2)}' | awk -v lex=$lexicon 'BEGIN{while((getline<lex) >0){ seen[$1]=1; } }
   {for(n=1; n<=NF;n++) {  if (seen[$n]) { printf("%s ", $n); } else {printf("<UNK> ");} } printf("\n");}' \
   > $cleantext || exit 1;
@@ -47,9 +48,8 @@ cat $cleantext | awk '{for(n=2;n<=NF;n++) print $n; }' | \
 cat $dir/unigram.counts  | awk '{print $2}' | get_word_map.pl "<s>" "</s>" "<UNK>" > $dir/word_map \
    || exit 1;
 
-# note: ignore 1st field of train.txt, it's the utterance-id.
 cat $cleantext | awk -v wmap=$dir/word_map 'BEGIN{while((getline<wmap)>0)map[$1]=$2;}
-  { for(n=2;n<=NF;n++) { printf map[$n]; if(n<NF){ printf " "; } else { print ""; }}}' | gzip -c >$dir/train.gz \
+  { for(n=1;n<=NF;n++) { printf map[$n]; if(n<NF){ printf " "; } else { print ""; }}}' | gzip -c >$dir/train.gz \
    || exit 1;
 
 train_lm.sh --arpa --lmtype 3gram-mincount $dir || exit 1;